2014

Постійне посилання на фонд

https://eir.nuos.edu.ua/handle/123456789/3337

Переглянути

Перегляд 2014 за Автор "Melnik, A. V."

Зараз показуємо 1 - 2 з 2
  • Ескіз
    Документ
    Probability analysis of risk-contributing factors in organizational tasks of ship repair
    (2014) Kovalenko, I. I.; Коваленко, И. И.; Shved, A. V.; Швед, А. В.; Melnik, A. V.; Мельник, А. В.
    The ship repair process is a complex technological process affected by rather large number of risk-contributing factors: the cost of repairs, repair durations, quality of repair, the presence of the necessary production facilities and personnel, etc. It highlights the prob-lem of analysis, prediction and forecasting of the impact of these factors in order to solve effectively a number of organizational measures which precede the implementation of the ship repair works. For the account of a variety of factors which affect the course of the ship repair process the automated information systems which contain the developed databases are developed at the enterprises. Because of the versatility the existing information systems of the dockyards management can not display the specifics of the ship repair and are mainly focused on the management of the financial costs at the enterprise. A common drawback of such systems is the lack of decision support modules in them which limits their function in the production management. The aim of this article is to consider the possibility of ap-pliance of the model to support the decision making under the conditions of risk which are presented by the probability trees at which the tasks of probabilistic inference and the devel-opment of a number of illustrative examples are solved. Preliminary, each risk-contributing factor which represents a system of random events is graphically displayed in the form of the distribution tree. Each branch of the distribution tree displays a single random event and its probability to be fulfilled. The combination of such trees received by their joint leads to the probabilities tree. Each node (the top) of such tree is connected with one complete sys-tem of random events. Each event and the probability of its fulfillment are displayed by the tree branch which comes from the corresponding node. Each path in the tree from the root node to the final position shows one of the possible combinations of events which are called a script. This approach can be successfully applied for the analysis of various organizational and technical problems of the ship repair under the conditions of uncertainty.
  • Ескіз
    Документ
    Представление прецедентов в базе знаний системы поддержки принятия решений при диагностике портальных кранов на основе теории грубых множеств
    (2014) Kovalenko, I. I.; Коваленко, И. И.; Melnik, A. V.; Мельник, А. В.
    Decision Support System (DSS) is a computer automated system aimed of the formation of recommendations to the decision maker in difficult conditions of uncertainties. The main element of the DSS is a knowledge base (KB) built on the basis of the precedents reasoning technique and on the efficient use of existing accumulated experience presented as a precedents base. To solve the problem of compact representation of a precedent in the for-mation of the BP for the first time there is provided to use the rough sets theory (RST). This article discusses the RST conceptual foundations, presents its basic definitions and terms, considers the basic operational states of portal crane (PC), their diagnostic criteria, which define the technical position of the cranes. Using the example of the BP formation for the DSS for the diagnosis of the portal crane (PC) the procedures of the elementary and fun-damental categories formation, equivalence classes based on the diagnostic parameters of the PC are demonstrated. Similarly, the procedure for determination of the upper and lower approximation, the boundary region of the target (rough) set and the assessment of accuracy and roughness of such sets are considered. This approach allows to perform a kind of "inac-curate" classification, which in practice may appear more real than the inability to perform the accurate classification.